Question Speculation: RDNA3 + CDNA2 Architectures Thread

[uzzi38]

Man I have been dying to make this one for a while now.

First rumours for RDNA3 are here so new thread time!

Just going to start off with this one for now: kopite7kimi on Twitter: "@VideoCardz Ah, I mean a simple mcm design with 10240 cores is not enough. Because the lift from RDNA2 to RDNA3 is much bigger than from RDNA1 to RDNA2. We should expect many big improvements in GFX11. 🤔" / Twitter

 

[Heartbreaker]

Maybe this is an alien world to you, but a LOT more individuals worldwide are going to be, if not already, experiencing financial problems, as in not being able to afford their dreams. They are the ones that will budget aggressively.

If I had the money, I would really, really want X, Y, Z, but I guess I have to settle for this. Better than nothing after all.

When people are struggling to afford necessities, lowering the price on luxuries, doesn't make them necessities. GPUs are not necessities.

I'm closer to the bottom than most here. I won't be buying a GPU at all, and a sale won't change that. When people are worried about job loss, food/shelter/fuel, then a cheaper GPU, is not going to convince them that's a good time to buy a new GPU.

 

[Timorous]

AMD will charge as much as they can possibly get away with to maximise profits, if they didn't their investors would not be happy. The only way we get cheap AMD cards is if they lack features and/or performance in key areas.

The right pricing model can set AMD up to start gaining market share in the GPU space the same way they did vs Intel in the CPU space. That leads to long term profits. Based on the rumoured info about RDNA3 it seems AMD could do that while maintaining or increasing their GPU margins which makes it a win win.

 

[maddie]

When people are struggling to afford necessities, lowering the price on luxuries, doesn't make them necessities. GPUs are not necessities.

I'm closer to the bottom than most here. I won't be buying a GPU at all, and a sale won't change that. When people are worried about job loss, food/shelter/fuel, then a cheaper GPU, is not going to convince them that's a good time to buy a new GPU.

It's not a binary decision, there are levels of struggling. Why this rigid thinking?

Individuals that bought higher end, slide down the price range. This applies to all purchases. Do we stop going to the movies or do we go less? Do we stop eating out or just at cheaper restaurants? Expand this to all aspects of life. Those already at, or close to the bottom financially will do completely without many things they previously had, as essentials consume their budgets. Those above will have less for discretionary spending and bargain hunt increasingly.

As an aside, and I'm NOT comparing now to then. Hollywood did well in the great depression. Why that happened, might have a lot to do with this discussion, modern gaming and the tech for it.

 

[Heartbreaker]

It's not a binary decision, there are levels of struggling. Why this rigid thinking?

There is very much two different camps. The Wealthy and secure who are unaffected, except for the current growth rate of the stocks and bonds, and those less secure concerned about necessities potentially being an issue.

For many/most of the less secure it is binary. Which is why many sales collapse. Recession sales don't drop, just because of current loss of income, but because of fears of job loss, so preemptive belt tightening back to necessities, and GPUs are FAR from necessities.

OTOH, For people with well above average income, and no fear of job loss, there is little impact from a recession and no need to curtail purchases.

So you end up in a situation where a 4090 will sell out at $1600, while low end $200 card sales slow dramatically regardless of price cuts.

You sound like you are just fantasizing about the personal benefit this recession might bring you in the form of your fantasy GPU price war.

 

[Kaluan]

AMD will charge as much as they can possibly get away with to maximise profits, if they didn't their investors would not be happy. The only way we get cheap AMD cards is if they lack features and/or performance in key areas.

It's not that simple. If AMD prices RDNA3 very aggressively against Lovelace and they have volumes and in 6 months they gain considerable mindshare and marketshare for the Radeon brand/products, I think their investors would be very happy. Seeing something you invested in start being very competitive in a segment they previously were meh (consumer/client GPU) and technically increasing revenue potential, is the next best thing long term investors would like to see, behind immediate ROI ofc.

 

[Heartbreaker]

Just because the current asking price on Amazon is $3000, doesn't mean that 4090s sold and contributing to the spot on the list were all sold at that price.

Good point, Amazon has multiple sellers for products. When the normally priced ones sell out, the scalper stores remain, hopefully most people aren't giving the scalper stores business.

Right now, all the normal priced online cards are gone.

 

[biostud]

One thing to keep in mind is that all of these companies made commitments to TSMC. They have to take certain number of wafers.

So selling small number of cards at high price does not necessarily increase margins, because the company already ordered (and has to pay for) wafers for large number of cards.

True, and while orders probably is down in all categories, then AMD has the possibility to switch between 6/7nm and 5nm lines and either GPUs or CPUs.

 

[Leeea]

It's not that simple. If AMD prices RDNA3 very aggressively against Lovelace and they have volumes and in 6 months they gain considerable mindshare and marketshare for the Radeon brand/products, I think their investors would be very happy. Seeing something you invested in start being very competitive in a segment they previously were meh (consumer/client GPU) and technically increasing revenue potential, is the next best thing long term investors would like to see, behind immediate ROI ofc.

I disagree.

AMD could hand out rx7900xts for free, and nvidia would still retain 80% of the market.


If it is free it must be defective, right?
If it is cheaper it must be inferior, right?

 

[gdansk]

I disagree.

AMD could hand out rx7900xts for free, and nvidia would still retain 80% of the market.


If it is free it must be defective, right?
If it is cheaper it must be inferior, right?

There's some truth to this line of thinking but you are exaggerating. People do want aggressive Radeon prices so they can buy cheaper GeForce.

 

[fleshconsumed]

I disagree.

AMD could hand out rx7900xts for free, and nvidia would still retain 80% of the market.


If it is free it must be defective, right?
If it is cheaper it must be inferior, right?

If I recall correctly ATI did get 50% market share with 3870/4870/5870 cards, but it took them 3 generations of aggressively priced cards that performed the same as nvidia to achieve it. It is a tough battle, but it is not impossible, it will take time and consistently delivering equivalent product at better prices or better product at same prices. The biggest hurdle is going to be CUDA for which AMD has no answer.

 

[Joe NYC]

True, and while orders probably is down in all categories, then AMD has the possibility to switch between 6/7nm and 5nm lines and either GPUs or CPUs.

After the big drop on client PC side, I think AMD has too much capacity on all of the nodes.

 

[maddie]

After the big drop on client PC side, I think AMD has too much capacity on all of the nodes.

We'll see soon if that quarterly drop was a one off. If it wasn't, ...............

 

[desrever]

If I recall correctly ATI did get 50% market share with 3870/4870/5870 cards, but it took them 3 generations of aggressively priced cards that performed the same as nvidia to achieve it. It is a tough battle, but it is not impossible, it will take time and consistently delivering equivalent product at better prices or better product at same prices. The biggest hurdle is going to be CUDA for which AMD has no answer.

That was 3 generation of Nvidia completely throwing it for AMD to gain that and it all evaporated in like 1 gen after Nvidia was competitive again.

 

[maddie]

There is very much two different camps. The Wealthy and secure who are unaffected, except for the current growth rate of the stocks and bonds, and those less secure concerned about necessities potentially being an issue.

For many/most of the less secure it is binary. Which is why many sales collapse. Recession sales don't drop, just because of current loss of income, but because of fears of job loss, so preemptive belt tightening back to necessities, and GPUs are FAR from necessities.

OTOH, For people with well above average income, and no fear of job loss, there is little impact from a recession and no need to curtail purchases.

So you end up in a situation where a 4090 will sell out at $1600, while low end $200 card sales slow dramatically regardless of price cuts.

You sound like you are just fantasizing about the personal benefit this recession might bring you in the form of your fantasy GPU price war.

I guess we are in a computer forum, and binary is the rule. Maybe we should advance this to the new quantum computing model.

 

[Joe NYC]

We'll see soon if that quarterly drop was a one off. If it wasn't, ...............

On another forum we tried to connect the dots, and the biggest culprit seems to be the inventory of PCs in the channel. The inventory level was 2 months greater than pre-pandemic.

And, it seems that all the parties involved wants to go back to that lower level of inventories.

This correlates very well with which parts of AMD revenue were hit (Client CPU). Other categories were not hit very much or at all.

From this, it would seem that this might be mostly one off...

But the dire economic situation in Europe and the US will have some negative effect across the board.

 

[biostud]

After the big drop on client PC side, I think AMD has too much capacity on all of the nodes.

You are probably right. But it makes you wonder about the total capacity of TSMC and over what time horizon those wafers has to be delivered. With both Intel and nvidia moving orders to TSMC, I doubt they have empty production queues in the foreseeable future.

 

[moinmoin]

The biggest hurdle is going to be CUDA for which AMD has no answer.

I disagree. Windows PC gamers seem to love Nvidia no matter what. Same can't be said about supercomputer users so there is a broad effort to move away from the current Nvidia controlled ecosystem, with AMD just one company in that effort. So the latter is both a "smaller" hurdle and a higher margin business.

 

[GodisanAtheist]

If I recall correctly ATI did get 50% market share with 3870/4870/5870 cards, but it took them 3 generations of aggressively priced cards that performed the same as nvidia to achieve it. It is a tough battle, but it is not impossible, it will take time and consistently delivering equivalent product at better prices or better product at same prices. The biggest hurdle is going to be CUDA for which AMD has no answer.

-AMD has toyed with 50% market share but the market has been a 60/40 split for a long time after the 5000 series and has collapsed all the way back to an 80/20 split as of 2022.

AMD ain't hunting market share, it's too brutal to gain and way too easy to lose.

Here is a nice 18 year market share recap video: short and easy to watch:

 

[biostud]

I disagree. Windows PC gamers seem to love Nvidia no matter what. Same can't be said about supercomputer users so there is a broad effort to move away from the current Nvidia controlled ecosystem, with AMD just one company in that effort. So the latter is both a "smaller" hurdle and a higher margin business.

I've had both AMD/ATI and nvidia cards, but for last couple of generations the price/performance has been equal or close to and with nvidia you got RTX and DLSS. Sure you got some extra memory on AMD, but currently it hasn't been that important. So for most you would get a better product from nvidia.
Now where a RX6900 is cheaper than a 3080 12GB, and FSR2.0 exists I think it is a better deal, unless you have a specific wish for raytracing. But it is too little too late. Hopefully they will fare better with RX7XXX.

 

[GodisanAtheist]

AMD has put a lot of shoulder into improving their DX11/OGL performance and reaching feature parity with NV, and that cannot be understated. You can see the effect additional margins and income has had on their efforts: they can put more money toward the "nice to haves", which in turn make them more competitive with NV, which in turn gets them more market share.

Selling cheap to manufacture cards at firesale prices that can only game and do nothing else doesn't get you real tractable market share, making money to make good products does, and that's why margins matter so much.

 

[Kaluan]

I disagree.

AMD could hand out rx7900xts for free, and nvidia would still retain 80% of the market.


If it is free it must be defective, right?
If it is cheaper it must be inferior, right?

...except nVidia is not God and neither is their mindshare bulletproof. Hell they've been getting massive bad press lately, from ridiculous pricing, tiering, fake SKUs, fake frames (DLSS should have just been called DLSS Ultra Performance, considering the fake performance you get and that image quality is eventhan the Performance preset on DLSS2, even at 4K) and last but not least, the fake/artificial performance gap they created by giving reviewers a early v522 driver in the form of the 4090-exclusive v521, while the older cards were reviewed on v517 or older.

4090 being down to just 40% faster than 3090 Ti, from 47% at 1440p and just 60% from 65% at 4K certainly made the 4090 look artificially better than 3000 series. nVidia spares no expense in researching new ways to screw over the consumer, doesn't it. 😂

Anyway, not to mention that Radeon had 50% market share before as well as the fact that Intel is a thing in the GPU scene rn (for how long, no one knows tho...)

Oh, there's also a thing called the deep recession we're heading into.

nVidia can only miscalculate so much...

 

[Kaluan]

Selling cheap to manufacture cards at firesale prices that can only game and do nothing else doesn't get you real tractable market share, making money to make good products does, and that's why margins matter so much.

Yeah I'm sure making the fat cat executives and shareholders get even more insane bonuses and fortunes hasn't got anything to do with it.

Please.

 

[Aapje]

This best seller list seems to be incorrect. Why would 4090 at $2999 be #2 on the top seller list?

This just reflects that these are short time sales, which makes then rather pointless. Of course a new GPU released just now is going to dominate the recent sales. But the 4090 is not going to be the 2nd most sold GPU.

 

[Vattila]

The biggest hurdle is going to be CUDA for which AMD has no answer.

AMD has a clear alternative to CUDA: HIP — a "CUDA dialect" serving as a programming model for their open-source ROCm framework, while supporting CUDA as an alternative backend. Intel and the wider industry have an alternative to CUDA as well: Khronos SYCL — a programming model based on ISO C++, aiming to be subsumed by the official ISO C++ standard at some point.

SYCL is conceptually pretty close to CUDA/HIP, except for accelerator code being written in pure C++ (no proprietary language extensions). A SYCL implementation already exists using HIP as a backend: hipSYCL. Also, Intel's C++ dialect — Data Parallel C++ (DPC++) — is based on SYCL and serves as a programming model for their open-source oneAPI framework. USA national laboratories have contracted Codeplay (now a subsidiary of Intel) to add support for AMD and Nvidia hardware in DPC++, thereby allowing SYCL/DPC++ to be used to write portable code across all these platforms.

SYCL is seeing rapid adoption in the supercomputing space, with AMD having delivered the first exascale computer (Frontier) this year and being on-track to deliver an even bigger one next year (El Capitan). These machines have a heavy emphasis on GPGPU and AI compute and will be programmed with no CUDA in sight (SYCL, HIP, OpenMP, etc.). AMD also won the biggest supercomputer in Europe (LUMI in Finland), which is based on the same hardware technologies and programming models. Intel will bring another exascale supercomputer online next year as well (Aurora), which will also use SYCL (DPC++) as one of the primary programming models.

More about SYCL, HIP and CUDA here:

Speculation: SYCL will replace CUDA

I've had my eye on SYCL for some time now. I don't think it is widely recognised that this is going to be the standard for heterogeneous programming models going forward (eventually to be merged into the ISO C++ standard, it seems). If you read any discussion on programming, tech and investor...

forums.anandtech.com

PS. I see a lot of discussion about the CUDA software moat on forums, without any mention of HIP, SYCL and DPC++ (oneAPI) whatsoever, which tells me there is a lot of ignorance when it comes to the industry's preferred direction (open standards) and the CUDA alternatives already gaining ground (SYCL and HIP). That said, AMD's lack of support for ROCm on Windows is a sore point, especially due to missing (or poor) acceleration support in some application areas (AI and content creation, in particular).

 

[insertcarehere]

It's not that simple. If AMD prices RDNA3 very aggressively against Lovelace and they have volumes and in 6 months they gain considerable mindshare and marketshare for the Radeon brand/products, I think their investors would be very happy. Seeing something you invested in start being very competitive in a segment they previously were meh (consumer/client GPU) and technically increasing revenue potential, is the next best thing long term investors would like to see, behind immediate ROI ofc.

AMD GPUs, by-and-large, need to compete with AMD CPUs for TSMC 5nm wafers.

Navi 31 + 32 all take up 200++mm^2 of N5 die space + are 5-7 chiplet package designs. From a packaging perspective this puts them closer to Zen 4 EPYC than anything else AMD sells. From that perspective it is very difficult for AMD and its investors to justify selling RDNA 3 for cheap when the opp cost here is less high-margin Zen 4 workstations.